10 ChatGPT Photo Editing Prompts That Keep the Face the Same

ChatGPT Photo Editing: 10 prompts that won't wreck the face

Last Verified: September 2026 — Quick-scan cheat sheet: copy a prompt, fill the brackets, go

Type “fix this photo” into ChatGPT and you’ll often get a face that looks airbrushed into someone else. That’s a missing instruction, not a limit of the model. Anything you don’t explicitly protect is fair game, skin included. Every prompt below has two parts: what to change, and what to leave alone.

TL;DR

  • Upload the photo first, then paste the prompt that matches your edit.
  • Every prompt ends with a “keep unchanged” clause. Don’t delete it.
  • If something drifts, name the drift in your next message instead of rewriting the whole prompt.
On this page (9 sections)
  1. Cheat Sheet at a Glance
  2. The Prompts
  3. Fixing Drift: The Follow-Up Library
  4. The Side-by-Side Check
  5. Combining Edits
  6. Before You Share
  7. Common Mistakes to Avoid
  8. Q&A
  9. Related Guides

Cheat Sheet at a Glance

EditPrompt #
Everyday retouch1
Remove an object or person2
Swap the background3
Sharpen a blurry shot4
Change the outfit5
Expand the canvas6
Change time of day7
Open closed eyes8
Soften face wrinkles9
Color grade10

The Prompts

1. Professional retouch

Use it for: Portraits that need cleanup, not a makeover.

Retouch this photo professionally: even out skin tone, reduce blemishes, sharpen focus, and balance lighting and color. Do not change the person's face shape, features, proportions, hairstyle, or expression. Keep the result looking like a real, unedited photo of the same person.

Sample run · Gemini 3.1 Flash Image (Nano Banana 2), October 2026

Input we used: Prompt 1 exactly as written, run on a real photo with visible acne (Pexels, free license).

Before: original photo
Before: original photo
After prompt 1: retouched
After prompt 1: retouched

The breakouts on both cheeks are gone, but her eyes, brows, nose, and mouth line up with the original. The skin did get a little smoother overall; if yours looks waxy, send the “skin looks plastic” follow-up below.

2. Remove an unwanted object or person

Use it for: Photobombers and clutter.

Remove [describe the object or person] from this photo and rebuild what's behind it. Match the existing lighting, shadows, texture, and perspective. Leave everything else unchanged.

Sample run · Gemini 3.1 Flash Image (Nano Banana 2), October 2026

Input we used: Prompt 2 with the slot filled as “the parked cars and the street lamp on the right”, run on a real CC0 public-domain photo from StockSnap (stocksnap.io), the same photo for every run below.

Before: original CC0 photo
Before: original CC0 photo
After prompt 2: cars and street lamp removed
After prompt 2: cars and street lamp removed

The cars and lamp are gone and the street was rebuilt with houses and a grass verge. Her face, hair, and top did not move, which is what “leave everything else unchanged” is for.

3. Swap the background

Use it for: New setting without a reshoot.

Replace the background with [describe the new background] and keep the subject exactly as they are: same pose, face, hair, clothing, and proportions. Adjust the lighting on the subject so they look photographed in the new setting.

Sample run · Gemini 3.1 Flash Image (Nano Banana 2), October 2026

Input we used: Prompt 3 with the background slot filled as “a quiet beach at sunset”, same photo.

After prompt 3: background swapped to a sunset beach
After prompt 3: background swapped to a sunset beach

The face held and the light now comes from the low sun, but look at the pose: she is leaning on a post that was never there and her head turned. That is drift. Follow up with “Keep the original pose and head angle exactly” and run it again.

4. Sharpen without losing realism

Use it for: Blurry or poorly lit shots.

Sharpen and clean up this photo: reduce blur and correct exposure while keeping the composition, subject, and every detail exactly as they are. Avoid any over-processed look.

Sample run · Gemini 3.1 Flash Image (Nano Banana 2), October 2026

Input we used: Prompt 4 exactly as written, run on a dim black-and-white CC0 portrait from StockSnap, then one follow-up.

Before: dark, low-detail original
Before: dark, low-detail original
After prompt 4: brighter, but colorized
After prompt 4: brighter, but colorized
After one follow-up: black and white again
After one follow-up: black and white again

The first pass fixed the exposure and detail but also added color to a black-and-white photo, which the prompt never asked for. One follow-up fixed it: “Colors were added. This photo was black and white: match the original black-and-white look exactly and keep the brighter exposure and sharper detail from this edit.” That is the drift library working: name the one thing that changed, keep the rest.

5. Change the outfit

Use it for: Trying a different look.

Change the outfit to [describe the new outfit], fitted naturally to the body. Keep the person's face, hair, pose, hands, skin tone, and background unchanged.

Sample run · Gemini 3.1 Flash Image (Nano Banana 2), October 2026

Input we used: Prompt 5 with the outfit slot filled as “a fitted navy blazer over a plain white t-shirt”, same photo.

After prompt 5: navy blazer, white t-shirt
After prompt 5: navy blazer, white t-shirt

Face, hair, and background survived. The model added a raised hand on the railing even though the prompt lists hands and pose as off-limits; outfit swaps are where this happens most. Crop it out or re-run with the hands follow-up below.

6. Expand the canvas

Use it for: Tight crop to wider shot.

Extend this photo to a [aspect ratio] canvas by continuing the existing background naturally in every direction. Do not alter anything currently inside the frame.

Sample run · Gemini 3.1 Flash Image (Nano Banana 2), October 2026

Input we used: Prompt 6 with the slot filled as “16:9”, run on another CC0 shot of the same model.

Before: original 3:2 photo
Before: original 3:2 photo
After prompt 6: extended to 16:9
After prompt 6: extended to 16:9

The model invented a house on the left, more garden, and more street on the right, and kept her, the fence, and the light exactly where they were. Check the seams: the new fence posts should match the old.

7. Shift the time of day

Use it for: Flat afternoon to golden hour.

Change the lighting in this photo from [current time of day] to [golden hour / blue hour / night], adjusting the sky, shadows, and color temperature to match, without changing the subject or composition.

Sample run · Gemini 3.1 Flash Image (Nano Banana 2), October 2026

Input we used: Prompt 7 with the slots filled as “bright afternoon” and “golden hour”, same photo.

After prompt 7: afternoon shifted to golden hour
After prompt 7: afternoon shifted to golden hour

The cleanest result of the set: sky, shadows, and skin all warmed together, and nothing in the frame moved.

8. Open closed or squinting eyes

Use it for: The one blink in a group shot.

Fix the person's closed or squinting eyes so they appear naturally open, matching their gaze direction and expression. Keep every other part of the face and photo unchanged.

Sample run · Gemini 3.1 Flash Image (Nano Banana 2), October 2026

Input we used: Prompt 8 exactly as written, run on a real photo (Unsplash, free license), cropped to the upper body.

Before: squinting into the sun
Before: squinting into the sun
After prompt 8: eyes open
After prompt 8: eyes open
Before (second run, StockSnap CC0): squinting
Before (second run, StockSnap CC0): squinting
After prompt 8: eyes open, mouth drifted
After prompt 8: eyes open, mouth drifted

Both eyes are open now and her smile, freckles, bangs, and the sunglasses in her hand are unchanged. Note the gaze: the model had to guess where she was looking, so check that the eyes point where you want. The second run shows the usual drift: his eyes opened, but his half-open mouth closed and his gaze moved. Send “Keep the mouth and gaze exactly as in the original” to pull it back.

9. Soften face wrinkles without de-aging

Use it for: Portraits where the lines look harsher than they do in person.

Soften the wrinkles and fine lines on [the person's] face by about half while keeping their age, face shape, expression, and natural skin texture. They should look rested, not younger or airbrushed. Leave everything else in the photo unchanged.

Sample run · Gemini 3.1 Flash Image (Nano Banana 2), October 2026

Input we used: Prompt 9 with the slot filled as “the older woman’s”, run on a CC0 photo from StockSnap.

Before: original (face crop)
Before: original (face crop)
After prompt 9: barely softer
After prompt 9: barely softer
After one follow-up: clearly softer
After one follow-up: clearly softer

“About half” turned out too cautious: the first pass barely changed her. One follow-up made the difference visible: “Go further: make her wrinkles and fine lines clearly softer, especially on the forehead, around the eyes, and on the cheeks. She should still look like the same woman with real skin texture, not airbrushed or decades younger.” Cheeks and forehead are smoother now, and she still looks her age. Crops shown so the change is visible; the woman beside her was left alone.

10. Apply a cinematic color grade

Use it for: A finished, intentional look.

Apply a cinematic color grade: balance highlights and shadows, and adjust contrast and saturation for a polished, consistent tone. Keep skin tones natural and leave the subject and composition unchanged.

Sample run · Gemini 3.1 Flash Image (Nano Banana 2), October 2026

Input we used: Prompt 10 exactly as written, same photo.

After prompt 10: cinematic grade
After prompt 10: cinematic grade

Deeper contrast and richer color, skin still reads as skin. Compare it with the original above: the composition is identical, only the tone changed.

Portrait on an easel with a glass shield over the face while sliders adjust the lighting
Lock the face first, then change everything around it.

Fixing Drift: The Follow-Up Library

Even with a good prompt, image edits sometimes drift. The fix is a short, specific follow-up, not a new prompt from scratch. Keep the original photo open next to the result and compare, then send the matching line:

What driftedSend this
Skin looks plasticSkin looks too smooth. Restore natural skin texture and pores; keep the rest of this edit.
Face looks like someone elseThe face changed. Match the original face exactly: same eye shape, nose, jaw, and smile.
Hands look wrongFix the hands: five fingers each, natural proportions, same pose as the original.
Text or logos got garbledRestore the text on the shirt/sign exactly as it appears in the original photo.
Colors shiftedColors shifted warmer. Match the original white balance except where I asked for changes.
Background warpedThe straight lines in the background bent. Keep walls, doors, and horizon straight.

One drift per follow-up works better than listing five problems at once. Each extra instruction is another thing the edit has to balance.

The Side-by-Side Check

Before you save or share an edit, check the areas image models most often change without being asked:

  • Eyes and teeth: shape, color, and whether both eyes point the same way.
  • Hands and jewelry: finger count, rings, watches, and anything someone holds.
  • Text: shirts, signs, and labels are often redrawn as near-miss spellings.
  • Edges: hair against the background, glasses frames, and ears.
  • Background people: faces in the background can change even when the subject doesn’t.

Zoom to 100% for this. Many drifts are invisible at phone-screen size and obvious on a print or large monitor.

Combining Edits

It’s tempting to ask for a retouch, a background swap, and a color grade in one prompt. Don’t. Each edit regenerates the image, and stacking three jobs gives the model three reasons to change the face. Do the edit that changes the most first (usually the background or outfit), check it, then retouch, then color grade last. If a later step breaks something, you still have the earlier version.

Portrait photo under a magnifying glass beside a shield and a paper plane
Check the face, the hands, and the small details before an edited photo goes anywhere.

Before You Share

Keep the unedited original. Some platforms and publications ask whether an image was edited with AI, and an original is the easiest way to show what changed. For photos of other people, get consent for both the edit and where it will be posted.

Common Mistakes to Avoid

  • Dropping the “keep unchanged” half. That’s what causes faces to drift and skin to over-smooth.
  • Rewriting the whole prompt after a bad result. Name the specific problem: “skin looks too smooth, keep natural texture”.
  • Editing photos of other people without consent. Get permission before editing and sharing someone else’s image.

Q&A

Does this work on ChatGPT Free?

Yes, with limits. OpenAI’s pricing page lists limited and slower image generation on Free, with more on paid plans.

Why does the face still change a little?

The edit regenerates the image, so small changes can happen. Compare side by side and name what drifted in a follow-up.

Comments

3 responses to “10 ChatGPT Photo Editing Prompts That Keep the Face the Same”

  1. […] complete instruction once: subject, style, lighting, and, for edits, what must stay unchanged. Our photo editing prompts and product photo template are built to land on the first or second […]

  2. […] asks the least of you. Image editing in particular is a daily-use feature for many people. See our photo editing prompts for how to keep edits under […]